This invention relates to a solid-state imaging device using a four-phase charge-coupled device (CCD) having four gate electrodes for a single photodiode and to a method of manufacturing the solid-state imaging device.
A CCD is a semiconductor device that has a structure where plural phase charge transfer electrodes are formed on a semiconductor channel region formed in a semiconductor substrate via a thin gate insulating film of a thickness ranging from 0.1 to 0.2 .mu.m, for example, and that transfers a charge signal by applying pulse voltages to the plural phase charge transfer electrodes, respectively, and changing the potential of the channel region under the electrodes.
FIG. 11 shows the structure of one cell of a conventional four-phase CCD. FIG. 11A is a top view of the cell. FIG. 11B is a sectional view taken along line XIB--XIB in FIG. 11A. FIG. 11C is a sectional view taken along line XIC--XIC of FIG. 11A.
As shown in FIG. 11A, a CCD solid-state imaging device is composed of a photodiode 11 constituting a light-receiving section (pixel section) and four phase charge transfer electrodes formed in the direction of, for example, column of the photodiode 11. On the photodiode 11, the charge transfer electrodes are not formed. Between photodiodes 11 and, for example, in a region in the direction of row, charge transfer electrodes are formed via a gate insulating film 2 above a semiconductor substrate 1, thereby forming a charge transfer region as shown in FIG. 11B, where its sectional view is shown. The charge transfer electrodes formed in a region in the direction of, for example, column between photodiodes 11 constitute an interconnection layer region for supplying voltage to the individual charge transfer electrodes in the charge transfer region as shown in FIG. 11C, where its sectional view is shown. Although not shown, on the regions except for the light-receiving section, a shading film is formed of a metal layer film made of, for example, Al.
The charge transfer electrodes shown in the figure are formed of three layers of polysilicon film: a first charge transfer electrode is made of a first-layer polysilicon film; a second and fourth-layer charge transfer electrodes are made of a second-layer polysilicon film, and a third charge transfer electrodes is made of a third-layer polysilicon film. The individual charge transfer electrodes are isolated from each other by an oxide film. The oxide film is formed by patterning the polysilicon film of each layer using, for example, lithography and etching techniques and by thereafter oxidizing the patterned polysilicon layer.
To form an oxide film for isolating the individual charge transfer electrodes from each other as described above, it is necessary to pattern each polysilicon film separately. This approach therefore requires mating margins for patterning in processing each layer, which makes it difficult to subminiaturize cells. Furthermore, to secure the mating margins, the area of a photodiode 11 is reduced, leading to the problem of lowering the sensitivity.
As shown in FIG. 11C, in the interconnection layer region, three layers of polysilicon films are stacked one on top of another, so the step height between the interconnection layer region and the photodiode region 11 is large. In forming a shading film, this prevents the shading film from being formed sufficiently over the step portion and therefore light may enter the regions other than the photodiode 11, causing an erroneous signal.
In contrast, there is a method of thickening a shading film so that the shading film may be formed sufficiently even at the step portion. With this method, however, when the shading film on the photodiode region 11 is etched away, it is difficult to etch the film away sufficiently. Because the shading film is formed thicker in the periphery of the photodiode region 11, the amount of light reaching the photodiode 11 is smaller, lowering the sensitivity of the solid-state imaging device. For this reason, it is desirable that the thickness of the shading film should not be made thicker.
Moreover, since it is necessary to form and process three layers of polysilicon film, the processes are long and complex.
As described above, with the conventional solid-state imaging device and method of manufacturing the device, since the charge transfer electrodes are composed of three layers of polysilicon films, the step height between a region where three layers of polysilicon films are stacked one on top of another, such as an interconnection layer region, and the photodiode region is large, which prevents a shading film from covering the step portion sufficiently, resulting in the problem of permitting light to enter the regions other than the photodiode, thus causing an erroneous signal.
Furthermore, because the three layers of polysilicon films must be patterned separately, it is necessary to secure mating margins for patterning, leading to the problem of making it difficult to subminiaturize cells.
In addition, because it is necessary to process the three layers of polysilicon films, this results in the problem of making the manufacturing processes long and complex.